Brightness-controlled enhancement for soil image based on conic curve

文献类型： 外文期刊

第一作者： Zeng, Shaohua

作者： Zeng, Shaohua;Wu, Weiran;Xia, Yan;Zeng, Shaohua;Wu, Weiran;Xia, Yan;Wang, Shuai;Liu, Guoyi

作者机构：

关键词： Image processing; Brightness-controlled enhancement; Soil image; Gray-level mapping; Conic curve

期刊名称：SIGNAL IMAGE AND VIDEO PROCESSING （ 影响因子：2.3； 五年影响因子：2.0 ）

ISSN： 1863-1703

年卷期： 2024 年 18 卷 2 期

页码：

收录情况： SCI

摘要： Soil images collected under different illumination conditions can be enhanced with a brightness-controlled enhancement algorithm, so that the result image is similar to the real image of the same soil photographed under specific natural light. It can improve the accuracy of soil species identification with machine vision because the image features of distinct soil species are described with a uniform brightness benchmark. To address it, a nonlinear mapping algorithm based on conic curve is proposed to realize the brightness-controlled enhancement of soil image. Firstly, a weighted quadric curve is derived in theory to reconstruct the brightness mapping according to the brightness relation between a pair of soil images with different brightness. Then, the brightness transformation range and the bending degree of brightness mapping curve are searched by the idea of binary search and gradual approximation to meet the given target brightness. The enhanced image is finally obtained by weighting and fusing the results of curve mapping in RGB space and HSV space. Extensive experiment results demonstrate that the proposed method is superior to comparison methods for brightness migration. Our algorithm can improve the accuracy of CNN models by up to 4.94%. Subjective evaluation shows that the effective range of soil image brightness enhancement with the proposed method is plus or minus 30 gray levels of brightness. The simulation result proves its effectiveness in brightness-controlled enhancement of soil image.

分类号：

  • 相关文献
作者其他论文 更多>>

微信小程序